Graph neural networks for hourly precipitation projections at the convection permitting scale with a novel hybrid imperfect framework

Today's article comes from the journal of Environmental Data Science. The authors are Blasone et al., from the University of Trieste, in Italy. In this paper they introduce a graph-based neural network that learns how to translate coarse atmospheric conditions into fine-scale estimates.

DOI: 10.1017/eds.2025.10022